Back to skill

Security audit

Filter

Security checks across malware telemetry and agentic risk

Overview

This skill saves user-created filter rules to a disclosed local JSON file and does not show hidden network, credential, or destructive behavior.

Install this if you want reusable filter criteria saved locally. Avoid putting secrets or highly sensitive personal terms into filter criteria, and periodically review or delete the local rules file if rules become outdated.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill declares a local config path under ~/.openclaw/workspace/memory/filter/rules.json and its stated behavior is to create and store reusable rules locally, which implies file read/write capability. When those capabilities exist without explicit permission declarations, users and enforcement layers may not have clear visibility into what local data the skill can access or modify. In this context the behavior appears aligned with the skill’s purpose, but the missing declaration still creates a real security and governance gap.

Session Persistence

Medium
Category
Rogue Agent
Content
---
name: Filter
description: Create and manage reusable filter rules for email, news, search results, and structured task streams. Designed to reduce noise and preserve signal in OpenClaw workflows.
version: 1.1.0
metadata:
  openclaw:
Confidence
78% confidence
Finding
Create and manage reusable filter rules for email, news, search results, and structured task streams. Designed to reduce noise and preserve signal in OpenClaw workflows. version: 1.1.0 metadata: ope

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.